THE PERFORMANCE OF AFFIRMABLE BUSINESS ENTERPRISES
6.5 RESEARCH FINDINGS
6.5.1 General
The research into completed contracts was conducted through a structured questionnaire in the following areas:
• Time frames;
• Mobilization of contract securities;
• Management;
• Labour;
• Safety; and
• Product quality.
Table 6.25 summarises the areas of research and lists the selected primary questions from the structured questionnaire, which were subjected to tests for statistical significance. The results of the application of the T-test to the groups of primary questions are also recorded in the table for comparative purposes.
TABLE 6.25: Areas where statistical differences were observed between ABE and non-ABE perfonnance
Area of .... earch X" p-value T -tee! significance
Tlmefrolme 0,008
Contractor started on time 0,111
Establish site quickly and efficiently 0,044
Contractor kept to programme 0,013
Completed contract within allowed time 0,135
Contractor worked extra hours 0,783
Contractor worked over weekends 0,104
Contractor submitted payment on time 0,005
Securities 0,225
Contractor provided required security 0,486 Difficulties in providing securities 0,027
Mobilised securities on time 0,353
Management factors 0,047
Number of full time employees 0,015
Managed spending and cash flow 0,293
Supplies problems resulting from non-payment 0,054 Difficulty obtaining credit from suppliers 0,089 Agreement to final quantities without dispute 0,117
Maintained drawing register 0,044
Maintained good relations with architects staff 0,340 Reacted DOsitively to variations by architect 0,041
Labour factors 0,876
Mobilised suitable personnel 0,474
Hired majority of labour locally 0,113
Hired labour for the contract only 0,665
Submitted employees for external training 0,276 Submitted employees for on-the-job training 0,174
Fonnal contracts with employees 0,023
Anv labour unrest 0,201
Safety 0,009
Any major accident
-
Frequent minor accidents 0,000·
Provided suitable first aid facilities 0,069 Maintained the first aid facilities on site 0,010
Published safety procedures 0,153
Provided safetv-clothing and equipment 0,076
Product quality 0,003
Understood the purpose of the contract
-
Maintained contract standard 0,099
Produced a higher quality standard 0,015
Removed any portion due to poor quality 0,381
Used substandard materials 0,024
Used the cheapest suppliers 0,242
-V.lue stKluld be dISregarded .s only one minor 9<X:ident was reported on.
It may be observed from Table 6.25 that at p<O,05 the observed differences in perfonnance between ABE and non-ABE contractors are statistically significant with respect to time frames, management factors, safety and quality control. The statistical differences are not Significant at this level in respect of the mobilization of securities and labour factors.
6.33
The results of a T-test on the overa11 performance of ABEs on all factors, derived from the full data set. is presented in Table 6.26.
TABLE 6.26: T-test results for overall performance scores of ABEs and non-ABEs on all factors
Number Mean OOg .... of
OVerall performance score T·value Significance +
of cases score Freedom
ABE
I.
238,94 -2,33 42 0,006"Non·ABE 25 294,00
..
p < 0,01, + two taIledThe T-test indicates that there is statistical significance in the overall performance between ABEs and non-ABEs across all of the factors recorded. The degree of significance (P=0,006) serves to disprove the null hypothesis Mthat no performance differences exist between ABEs and non-ABEs in general contracts in the price range R100 000 to R2 000 000·, and thereby confirms the sub-hypothesis.
6,5,2 Time Frames General
The author's research into aspects of contractor performance against time frames examined aspects of relevance to the management of the contract. As such, the research focused on aspects such as commencement and completion times, site establishment and adherence to programmes. In addition, the matter of submission of payment certificates is covered, which is relevant to management of cash flow for both the Department and the contractor.
Late contract start
The research probed concerns relating to the ability of contractors to furnish securities and accept hand over of the project site from the Department. The author proposes that records of performance in this first phase of a project illustrate important management capabilities of contractors.
The bivariate analysis of sample data indicates in Table 6.7 that delays in the start of contracts occur frequently in general contracts in the research price range. Almost 30% of contracts start late, and both ABE and non-ABE contractors are liable to start late. The bivariate analysis from the sample suggests that ABEs are more likely to start late, but the
>f
test does not attach statistical significance to this observation. The author contends, therefore, that the Department can anticipate that roughly a third of general contracts having prices below R2 000 000 will start late and that no difference in performance is expected between ABE and non-ABE contractors.The bivariate analysis suggest that 75% of projects which start late will start within 8 weeks of the planned start date, whilst almost 40% will start within 4 weeks of planned start date. The observed
trends in the data suggest that ABEs which start late tend to start later than non-ABEs. The data records that 25% of late starting ABEs start within 4 weeks of planned start date, whilst 60% of late starting non-ABEs made a start within this period. The level of response to this secondary question in this research questionnaire was not sufficient to conduct the
>f
test. As a result, no statistical significance can be attached to the observation. This aspect of contractor performance therefore remains open to further research to obtain a sample of sufficient size for the application of the more sophisticated techniques.Recorded difficulties encountered by ABEs, which caused delays, were; mobilisation of security, late completion of previous DPW jobs, lack of adequate human resources at the contractor's disposal, lack of adequate equipment on site (e.g. plant), and financial problems encountered by the contractors.
Reasons presented for the delay in contract starts by non-ABEs were: consultation between head office and regional office leading to late hand-over of site to the contractor, contractor not delivering material to the project site in time, cancellation of contract due to failure by contractor to commence or start within reasonable time thereafter. None were recorded as being due to difficulties in mobilising security.
A common problem to both ABE and non-ABE contractors has been the inability to commence on time due to commitments on other jobs awarded by the DPW. This suggests that these contractors are at times awarded jobs which result in workloads their businesses cannot handle. This does not benefit the APP objectives and neither does it benefit the contractors in any way to be penalised for finishing late.
Variable allowances for late start were found in the research. The management style of different project managers rather than set departmental procedures appears as a determining factor.
Reluctance of managers to cancel or re,tirect a contract once some progress in the appointment has been made may underly these differences.
The recorded data on factors which created late starts were not adequate for non-parametric testing and no statistical Significance can be attached to them. Results show tendencies to differences between ABE and non-ABE contractors, which will be better illuminated if additional research is carried out. The data suggest that ABEs tend to have endogenous business factors causing delays. The identified causes include financial delays and personnel or plant shortages. The identified causes for non-ABE delays are almost entirely exogenous to the enterprises, where the Department has been late due to contract changes or late site hand over. Further research to clarify any differences between ABEs and non-ABEs in this respect is recommended. The author asserts that better knowledge will assist the Department to focus APP adjustments on this part of the project cycle.
6.35
The research indicates that any programme of service delivery that is managed by the Department through contracts conforming to the research type and price range is most likely to experience start delays in 30% of projects in the programme. This wilt have impacts on management focus and effectiveness, and on cash flow planning for the Department. It will also impact on the Departments clients, whose planning for related activities or expenditure wilt be similarly affected. The author is not aware of research in public works programmes with simitar characteristics to this research that provides a benchmark for late start frequencies. Data has not been locally researched for contracts above R2 000 000 let by the Department. The author contends that the level of late starts that has been identified is high.
Site establishment
The research investigated the efficiency of site establishment by contractors once a site has been formally handed over in terms of the contract. The bivariate analysis of research data reported in Table 6.7 suggests that non-ABEs are more effective in this phase of a project. They are indicated to establish a site quickly and efficiently in 92% of researched projects. In contrast, ABEs were evaluated to be efficient in almost 70% of projects. The
>f
test indicates a p-value of 0,044, confirming that this observed difference in performance between the two types of contractor is statistically significant within the p<0,05 limit.The author concludes that once contracts are active ABEs will benefit to a greater extent than non- ABEs from any assistance with site establishment that may be given by the principal agent or project manager.
Contractor adherence to programme and timeous completion of contract
The bivariate analysis presented in Table 6.7 suggests that ABEs are more likely than non-ABEs to keep to programme by a small margin. The data indicate that 7S%of ABEs and 72% of non-ABEs in the sample maintained their project programme. The data suggest further that ABEs are likely to contain slippage more effectively than non-ABEs to within 4 weeks. Application of the
>f
test to this data confirmed statistical significance in the observed differences at p=0,013 level.Slow progress in certain stages of the project, faults identified in manufactured products, delays in commencing with the work, unjustifiable absence from site for lengthy periods at a time, and poor workmanship resulted in the observed ABE failures in keeping up with their programmes.
Matters reported which contributed to failure of non-ABEs to keep up with programmes and complete in time were variation to the work and subsequent waiting for the approval of the extended budget, the bankruptcy declaration of a contractor and liquidation during execution of a contractor, and increases in scheduled and non-scheduled rates.
Interpretation of the bivariate data for completion of contract on time, and working extra hours and weekends suggest that non-ABEs perform better (84% finishing on time compared with 75% of ABEs) and make up time lost in their programme more effectively than ABEs. Application of the
XZ
test doesnot indicate statistical significance in these observations. As recorded in the Table 6.8 not all of the cells in these tests are filled. and a larger sample may increase the strength of the observations.
The author concludes that the research suggests that ABEs may manage slippage better than non- ABEs but are less likely to overcome slippage and complete contracts on time. Non-ABEs may improve their project completion performance by working extra hours and weekends more frequently than ABEs.
Whilst the research remains inconclusive, it suggests, through comparison of rates of late start and late finish, that ABEs appear able to improve from 42% late start to 25% late finish. The data for non-ABEs suggest a less dramatic improvement in performance, from 20% late start to 16% late finish. The research data reported in Table 6.7 does not clearly indicate reasons for the relative differences in performance.
The research indicates that approximately 20% of all contracts of this type and price will finish late.
This will impact on programme planning and on Department and client cash flow estimates and budgets in a similar manner to that discussed above for late project starts. Because the rate of late completion is lower than that of late start (approximately 20% versus 30% of all contracts) the author reinforces here the recommendations for the introduction of flexible project start management. Any improvement in rates of late start that can be achieved through flexible management techniques will have more than proportional impact on the rate of late finishes due to the observed ability of contractors to reduce contract slippage.
The author's research has not revealed comparative research that can confirm or contradict these conclusions, or that can provide comparisons for the rates of late contract completions. The author recommends that the Department conduct further research in these aspects to confirm the observations and to determine whether there are additional management procedures that may be influenced to change overall performance of contractors in these aspects.
The author proposes that these findings extend the structural understanding of this sector by indicating that project time management problems are not confined only to ABE contractors. The Department and its clients will benefit from any reduction in the scale of late contract completions achieved in the research contact price range.
6.37
Submission of payment certificates
The research data collected is relevant for reviewing contractors' management of their cash flow, which is an essential element of sustaining and developing a viable enterprise. The bivariate data in Table 6.7 indicate that almost all non-ABEs can be expected to submit payment certificates timeously (96%
of the sample did so). In contrast over one third of ABEs (37%) are likely to be late with submission of payment certificates. The
>f
test confirms the statistical strength of the observed differences in performance to a probability of occurrence of 99,5% (p=O,005). The author has not tested the relationship between poor management of payment certificates and late completion by ABEs due to the small size of data, and the fact that some of the cells in this test were empty.Results of poor cash flow management are likely to be problems with the purchase of materials and payment of labour. Both of these may prejudice the survival or growth of an enterprise. Comments from participants in the research point to a causal relationship between use of quantity surveyors and timeous submission of payment certifICates. Respondents noted that those ABE contractors who did not use quantity surveyors to assist with payment certificate submission were irregular in submissions, suggesting that speCialised skills are required in this area. A considerable percentage of ABEs are reported to have relied on quantity surveyors' expertise to make timely submissions.
The author concludes from the bivariate data in Table 6.7 that a greater degree of problem performance is indicated for ABEs in this factor than for late completions (37% versus 25%), and that some ABEs are able to survive despite poor cash flow management. A major improvement in performance therefore appears possible through supporting the employment of quantity surveyors by ABEs. The results are clearly demonstrable to contractors from the data reported here. The author recommends consideration of this motivation by the Department and supports it by indicating that it will provide external input into ABE projects. As such contractors themselves do not need to acquire training or experience and measurable results will be achieved in a short time without attracting time and fee-based costs for contractors. The author anticipates that such action may improve the rate of timeous project completion by ABEs by virtue of improved cash management, and by improving enterprise viability This may have a knock on effect in improving timeous start rates, where contractors win consecutive tenders from the Department.
General data relationships
The results of the application of the T-test are presented in Table 6.10 and summarised in Table 6.25.
The level of significance of 0,008 confirms that the observed differences in performance between ABEs and non-ABEs in the data are statistically significant.
What was not tested in the research was whether or not the contract period which was set is realistic.
The fact that both ABEs and non-ABEs alike experienced late starts, suggests that research is required
to establish what a realistic contract period for contracts in this value range should be. Watermeyer and Band (1994) identified -contract duration-as being a SMME barrier to entry, but assigned a low value to it. The weighting assigned to this should be increased in the light of the current research.
8.5.3 Contract securitle. factors
In this section the author discusses the data presented in section 6.3.3.3. The issue of provision of project securities has been recognised by a number of researchers. The author designed part of the data collection to test the strength of the generally held concepts under the experience of the APP.
The provision of contract securities has always been regarded as being an SMME barrier to entry.
Waterrneyer and Band (1994) in their survey rate the provision of sureties as being the highest ranked barrier to entry. This barrier in their research was ranked higher than Mlack of financial resources· and Minability to obtain credit.·
The reporting of bivariate analysis in Table 6.1
°
suggests that ABEs are more likely to experience difficulties than non-ABEs in providing securities. Proportions of contractors who experienced difficulties are recorded as 32% and 4% respectively. The application of thex!
test indicated statistical significance in the observed differences, where p=0,0272.This difference between ABEs and non-ABEs does not appear to influence the overall performance of the two types of contractor. The bivariate analysis reported in Table 6.10, suggests that observed occurrences of provision of securities occur almost equally by both types of contractor, in approximately 90% of cases. The application of the
>f
test does not indicate statistical significance to the observed minor differences between contractor types. The author therefore concludes that levels of provision of securities are almost the same for the two contractor types.More importantly, the author concludes that both ABEs and non-ABEs are equally likely to provide securities on time, with about 16% of all contractors not being on time. This conclusion is derived from the observations indicated by the bivariate analysis and by the lack of statistical significance in the observed differences (p=0,353) indicated by the
x:
test. The Department can therefore anticipate delays in its scheduling of projects within a programme or projects up to this stage in the project cycle.Research data report that some problems have been experienced with continuity between projects, where contractors' guarantee capacity has been tied up in an existing, incomplete project. Some flexibility has been shown by the Department to resolve these occurrences; in particular by creating a retention reserve from cash flow. Whilst solving the problem in the short term, the impact of this practice on contractor cash flow viability needs further investigation.
2 Not all the celts in the test were populated, and the author recommends additional research to confirm this finding.
6.39
The author submits that the observations above advances the understanding of issues related to the provision of securities because comparative research has been done. The )(2 test suggest that ABEs are more likely to experience difficulties in providing securities. However, no differences between ABE and non-ABEs are observed in respect of actually providing the security on time. Thus ABEs appear to be able to provide securities, even through some may experience some difficulties. Ultimately, the performance between the two types of contractor does not differ. The T-test result for these factors confirms this. (A significance level of 0,225 is reported in Table 6.12 and summarised in Table 6.29.)
The research findings do not support the outcomes of the survey reported on by Watermeyer and Band (1994) which ranked the provision of sureties as being the most important barrier to new entrant SMMEs to overcome. The research was however conducted on a closed sample of contractors who had in fact been awarded contracts and any conclusions should as a result be viewed with caution.
The author suggests that the strategy of reducing the levels of surety between 2,5 and 5% from 10% in this value range of contracts may be the reason for this finding. The author accordingly proposes that the Watermeyer and Band rating for this item be down graded should the levels of sureties called for be in accordance with government's 10 Point Plan (MOF and MOPW, 1995).
6.5.4 Management Factors
The author researched general aspect of management, seeking to establish relationships to overall performance. Although the sample size does not support the exploration of relationships between factors, valid observations are derived.
The observed differences between ABEs and non-ABEs are shown to be statistically significant in 3 of the 8 management factors under the)(2 test, and generally by the T-test. They are thus considered to generally support the sub-hypothesis.
The comparative nature of the research has, however, provided a clearer picture of the comparative performance of the types of contractor. The three factors, on which the )(2 test revealed statistically significant relationships, are not considered to be the most critical of all of the eight factors researched. They are the size of the firm, maintenance of a drawings register and positive reaction to variations by the architect.
No statistical differences in performance bet'oNeen the types of contractor are identified by the )(2 test for management of spending of cash flow, supply problems resulting non-payment, difficulties obtaining credit from suppliers, or agreeing to final quantities without despute. The bivariate analysis suggests differences in ABE and non-ABE performance. The)(2 test. however, must be interpreted with caution.